李云龙二次元风格化！一键star、fork，你也可以生成这样的团长！

打滚卖萌求star求fork!

1.首先要找到自己要操作的视频素材，将视频的音频单独提取出来备用

我自己找的资源放在了codes/videos/liyunlong文件夹下，是李云龙名场面：
你咋不敢跟旅长干一架呢！→旅长我给你跪下了
名场面

2.话不多说，进入代码实操：

  • 安装基本环境 
    !pip install -r codes/PaddleGAN-develop/requirements.txt
  • 导入基本环境 
    import paddle 
import os 
import sys 
sys.path.insert(0,'codes/PaddleGAN-develop')
from ppgan.apps import AnimeGANPredictor

3.GAN它！

友情提示：此处最好使用GPU环境，cpu推理属实是有点点慢

使用paddlepaddle预训练好的animeGANv2模型对视频进行风格迁移：
from ppgan.apps import AnimeGANPredictor
import cv2

predictor = AnimeGANPredictor('',None,)
video_src = 'codes/videos/liyunlong/格式工厂混流 亮剑-03+亮剑-03+亮剑-04 00_00_23-.mp4'
video_ = cv2.VideoCapture(video_src)
video_name_ = os.path.basename(video_src)
total_frames = video_.get(cv2.CAP_PROP_FRAME_COUNT)
fps_ = video_.get(cv2.CAP_PROP_FPS)
print("video {}, fps:{}, total frames:{}...".format(video_name_, fps_, total_frames))
frame_count_ = 0
save_per_frames = 1
dst_dir = 'codes/videos/liyunlong/'
out_video = cv2.VideoWriter('{}/hayao_{}'.format(dst_dir, video_name_),
                                cv2.VideoWriter_fourcc(*'DIVX'), int(fps_),
                                (int(video_.get(3)), int(video_.get(4))))
print('now begin...')
while True:
    ret_, frame_ = video_.read()
    if not ret_:  # or len(fps_list_) == 0:
        print('end of video...')
        break
    result_frame = predictor.anime_image_only(frame_)
    if frame_count_ % save_per_frames == 0:
        out_video.write(result_frame)
    frame_count_ = frame_count_ + 1
    if frame_count_ % 100 == 0:
        print("{}/{} processed...".format(frame_count_, int(total_frames)), flush=False)

4.合并生成的视频和之前分离的音频：

!ffmpeg -i codes/videos/liyunlong/hayao_格式工厂混流 亮剑-03+亮剑-03+亮剑-04 00_00_23-.mp4 -i codes/videos/liyunlong/音频1.aac -c:v copy -c:a aac -strict experimental codes/videos/liyunlong/李云龙二次元化.mp4

这样就大功告成啦~~~
打滚卖萌求star、fork！


微信公众号：璇珠杂俎(也可搜索oukohou)，提供本站优质非技术博文～～ wechat_official_account


regards.

oukohou.